[PATCH] D29826: [SLP] General improvements of SLP vectorization process.

[Alexey Bataev via Phabricator via llvm-commits]

ABataev added inline comments.


================
Comment at: test/Transforms/SLPVectorizer/X86/horizontal.ll:834
+; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <8 x i32> [[BIN_RDX4]], i32 0
+; CHECK-NEXT:    [[ADD_6:%.*]] = add nsw i32 undef, [[ADD_5]]
+; CHECK-NEXT:    [[RES:%.*]] = call i32 @foobar(i32 [[TMP1]])
----------------
RKSimon wrote:
> It's a bit annoying that the old scalar reduction hasn't completely disappeared.
This is how SLPVectorizer currently works. There is an outstanding patch D29641 that fixes this (and other) problem(s).


https://reviews.llvm.org/D29826